China-based tech giant Baidu (NASDAQ: BAIDF) has unveiled the latest iteration of its generative artificial intelligence (AI) chatbot barely days after authorities rolled out security measures for AI firms.
During the launch, Baidu CEO Robin Li stated that Ernie 4.0 “stands at par with GPT-4 in terms of overall performance” following a series of live demonstrations. Li remarked that the new iteration of the chatbot demonstrated advanced functionalities across the verticals of understanding, generation, reasoning, and memory.
The CEO disclosed to attendees of the launch event that the chatbot now has the ability to understand and interpret “out-of-order statements, vague expressions, and implied meanings” through text inputs. Li added that the generation capabilities of the chatbot have received a jolt in the latest version, transforming “a single person into a marketing team.”
A series of tests into Ernie 4.0 reveals above-par functionalities in problem-solving abilities, coding, and performing audits akin to OpenAI’s ChatGPT-4, dubbed as the industry standard for generative AI chatbots.
“Even after five rounds of conversation and writing thousands of words, Ernie Bot can remember previously generated content, showcasing the remarkable memory capabilities of Ernie 4.0 foundation model,” read a statement on X (formerly Twitter).
The latest model has been trained extensively in Mandarin, allowing users to interact with it using conversational prompts. Baidu’s chatbot was also trained using “massive valuable, real-world human feedback” to push the frontiers of generative AI.
Under the new regulatory regime in China, AI developers are expected to conduct security evaluations before launching new AI products for the market. Following the new rules, Baidu appears to have complied with the regulatory requirements preceding a commercial release.
Baidu launched Ernie 3.0 in August to critical acclaim as the application soared to the top of the Chinese app stores within 24 hours of its launch. Since its launch, the chatbot has enjoyed mass adoption levels, finding applicability in generating texts, images, and videos, but faces stiff competition from other technology firms in the country.
AI products flood Chinese market
After China unveiled interim measures for AI, over 70 generative AI models were commercially launched within the first 30 days. Tencent launched its Hunyuan AI model, while Alibaba (NASDAQ: BABAF) rolled out its open-source AI model, leaving consumers with various options.
Despite the seemingly tough stance, Chinese authorities are keen on promoting the development of locally designed AI systems to rival Western-based products like Meta’s Llama 2 (NASDAQ: META), Google’s Bard (NASDAQ: GOOGL), and ChatGPT. Akin to handling the digital currency ecosystem, Chinese authorities are keen on bringing AI under control to protect users and stamp out bad actors.